Targeting Oncoproteins with a Positive Selection Assay for Protein Degraders

Cancer genome sequencing studies have identified frequent driver mutations that alter the function of transcription factors (TFs). However, most TFs, are classically “undruggable”. The discovery that IMiDs [immunomodulatory drugs e.g.Lenalidomide (LEN) and Pomalidomide (POM)] exert powerful antimyeloma effects by triggering the proteasomal degradation of the critical TFsIKZF1 and IKZF3has led togreat interest in the discovery of“degraders”in other cancers. Existing “down” assays to identify destabilizers (e.g.measuring the loss of the protein of interest [POI]) often have(1) poor signal/noise ratios and narrow dynamic ranges and (2) generate uninterestinghits e.g.drugs thatinhibit transcription or translation. We present a novel positive-selection “up” assay, compatible with high throughput screening, for identifying drugs or sgRNAs that destabilize a POI. We have used this platform to identify novel destabilizers of IKZF1.

Our assay is based on a previously published “suicide gene”, a variant of the nucleotide salvage gene deoxycytidine kinase (DCK). This variant (DCK*) has mutations that increase its specificity for the synthetic substrate 2-Bromovinyldeoxyuridine (BVdU). Cells that express DCK* are killed by BVdU. We made mammalian expression vectors co-expressing GFP and DCK*, IKZF1, or a DCK*-IKZF1 fusion protein as a single bicistronic mRNA and stably introduced each into 293T cells. In the absence of POM, cells expressing DCK* or DCK*-IKZF1 are killed by BVdU . However, in the presence of an IKZF1 destabilizer, e.g. POM, the DCK*-IKZF1 protein is degraded, making the cells resistant to BVdU. Cells expressing DCK* alone are unaffected by POM and serve as a specificity control for the assay (Panel A).

We did a pilot chemical “up” screen with a ~ 2000 bioactive compound library (that included LEN and POM) using 293T cells expressing DCK*-IKZF1. In parallel, we conducted a “down” screen, using 293T cells co-expressing an IKZF1-Firefly Luciferase (Fluc) fusion protein and Renilla luciferase (Rluc) from a single bicistronic mRNA. Compounds that decreased the ratio of Fluc/Rluc activity were scored as hits. As expected, LEN and POM scored in both assays,but there was considerably more noise in the down assay. Lastly, we used the DCK*-IKZF1 cells to screen a library of uncharacterized IMiD derivatives. The screen correctly identified a novel IMiD derivative (MI-2-61) and a known next-generation IMiD (Avadomide) with greater potency against IKZF1 than POM.

To identify novel degraders of IKZF1, we used the DCK*-IKZF1 cells and DCK* control cells to screena metabolic inhibitor/anticancer library of ~600 compounds. We identified Spautin-1 as a compound that rescues DCK*-IKZF1 cells, but not DCK* control cells from BVdU toxicity. Spautin-1 downregulates exogenous IKZF1 in 293T cells (Panels B and C) and endogenous IKZF1 in KMS11 myeloma cells. Spautin-1 reportedly suppresses autophagy through inhibition of USP10 and USP13. However, siRNA mediated knockdown of USP10 and USP10 neither altered IKZF1 protein levels, nor blocked downregulation of IKZF1 by Spautin-1. Moreover, Spautin-1 downregulated IKZF1 in 293FT cells in which autophagy was disabled by CRISPR/Cas9-mediated disruption of genes crucial to the autophagy pathway suggesting that the downregulation of IKZF1 by Spautin-1 occurs via a novel mechanism.

Unlike IMiDs, the downregulation of IKZF1 by Spautin-1 does not require the CRBN E3 ligase. However, it is blocked by inhibitors of the E1 ubiquitin activating enzyme or the proteasome, but not by neddylation inhibitors required for Cullin-dependent E3 ligases. These data suggest that Spautin-1 triggers the proteasomal degradation of IKZF1 using a non-Cullen E3 ligase. The downregulation of exogenous IKZF1 by Spautin-1 requires the N-terminus of IKZF1, but not the zinc finger domain (ZF2) targeted by IMiDs. Preliminary structure-activity relationship (SAR) studies identified both active and inactive Spautin-1 derivatives, suggesting that downregulation of IKZF1 by Spautin-1 reflects a specific protein binding event and that Spautin-1’s potency and specificity can be optimized further.

We are undertaking studies to identify the mechanism by which Spautin-1 degrades IKZF1 in the hope that it may represent a novel druggable pathway to therapeutically degrade IKZF1, and validate the use of this platform to discover degraders of “undruggable” targets in other cancers.